基于多需求分布的收益管理多舱位无约束估计模型  

MULTI-CLASS UNCONSTRAINING ESTIMATION MODELS BASED ON MULTI-DISTRIBUTION OF DEMAND IN REVENUE MANAGEMENT SYSTEMS

作　　者：郭鹏[1] 萧柏春[1,2] 李军[1] 

机构地区：[1]西南交通大学经济管理学院,成都610031 [2]美国长岛大学管理学院,纽约11548

出　　处：《系统科学与数学》2013年第10期1164-1177,共14页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金重大项目(71090402);教育部长江学者和创新团队发展计划(IRT0860)资助课题

摘　　要：相较于单舱位无约束估计方法,多舱位方法能有效避免收益管理系统对同航班舱位间需求的高估问题.同时,无约束需求数据的多分布假设更加符合收益管理实践中不同情况下的需求特征.为此,首先建立了基于正态、对数正态和伽玛分布的多舱位Spill模型;然后,使用数值算例说明了所提方法的可行性;最后,通过与单舱位方法比较,验证了所提多舱位Spill模型的有效性,表明了数据的离散程度是影响各分布假设下Spill模型无约束估计效果的主要因素.Compared with single-class unconstraining methods, multi-class meth- ods could efficiently avoid the problem of over-estimating the demand across several classes within the same flight in revenue management systems. Meanwhile, the as- sumptions of multi-distribution for the nominal unconstrained demand is more suit- able for the demand characteristics in different circumstances under the practice of revenue management. For these problems, the multi-class Spill models in which the nominal demand data is assumed to follow a Normal, Lognormal, and Gamma dis- tribution are developed. And then, numerical examples are given to illustrate the feasibility of the proposed methods. Finally, compared with the single-class methods, the results show the effectiveness of the proposed multi-class Spill models, and indi- cate that the volatility level of demand data is the main factor that influences the accuracy of unconstraining process of the Spill models based on different distributions of demand.

关 键 词：收益管理 无约束估计 Spill模型 多舱位 需求分布 

分 类 号：F270[经济管理—企业管理] F224[经济管理—国民经济]